Decreased insulin action, impaired insulin secretion and increased adiposity are important risk factors for development of type 2 diabetes. These risk factors are present even when individuals have both normal fasting and two hour glucose concentrations indicating a very early role for decreased insulin secretion in the development of type 2 diabetes. We examined more closely the relationship of insulin secretion (as measured by the acute insulin response) to plasma glucose concentrations. We found that in Native Americans of southwestern heritage that the decline in the acute insulin response was non-linear with increasing glucose while there was a linear decline in whites. However, in both groups, declines in the acute insulin response preceded currently established cut-offs for impaired glucose regulation. We have also found that the one hour glucose concentration, which correlates well with insulin secretion and insulin action, is just as good a predictor of the development of diabetes as the 2 hour glucose concentration, and predicts diabetic retinopathy In previous work we found that a bimodally expressed gene in muscle tissue, HLA-DRB1, was associated with lower risk for diabetes and higher insulin secretion in Native Americans of southwestern heritage. As this protective HLA haplotype system indicates involvement of autoimmunity in progression to type 2 diabetes in Native Americans of southwestern heritage, we performed a pilot study in individuals with and without the HLA-DRB1 locus who were also discordant for diabetes status, examining both immunoprofiling of over 9,000 potential proteins and T-cell receptor genes. From among those 9,000 proteins study we selected 80 antibodies with significantly increased concentrations in those with diabetes and tested these in a larger cohort of 45 individuals with diabetes and lacking HLA-DRB1 and 45 individuals with normal glucose regulation with the HLA-DRB1 allele. These individuals were matched for age and sex. Of the selected proteins, 11 again had higher signals in those with diabetes. One of the identified antibodies was to PPARG, a nuclear receptor involved in adipocyte differentiation and a target of anti-diabetic medications. We then tested these antibodies in a large case control study, and found that autoantibodies against prefoldin 2 (a subunit of prefoldin, a molecular chaperone which facilitates protein folding) were higher in subjects with diabetes compared to those with normal glucose regulation. Furthermore this effect was much stronger in those with younger onset (age <30 years) type 2 diabetes indicating that autoimmunity may play a greater role in disease progression in younger individuals. Auto-antibodies to PPARG were not associated with diabetes, but were associated with lack of HLA-DRB1 allele. We have measured prefoldin 2 antibodies in a case control study to determine if these predict onset of type 2 diabetes at age <30 years, these analyses are ongoing. Previous predictors of weight gain based on this study have included, higher respiratory quotient, higher insulin mediated glucose uptake, lower free T3, and relatively lower energy expenditure. Variability in energy expenditure is a mediator of weight change, and we have continued to evaluate factors related to metabolic rate. We had previously confirmed this association between lower energy expenditure relative to body size as a predictor of weight gain in a larger cohort with longer follow-up. Furthermore, we were able to demonstrate that lower energy expenditure also predicted gain in fat mass. Using resting energy expenditure measured at 5 and 10 years of age, we were also able to demonstrate that relatively lower resting energy expenditure at age 10 (but not at age 5) predicted greater increase in body mass index indicating an effect of metabolic rate on weight in late childhood. The thermic effect of food and the energy cost of arousal are overlapping and difficult to measure components of energy expenditure. By examining time point data from our metabolic chambers, we were able to estimate these components (which we termed awake fed thermogenesis (AFT)) which accounted for approximately 10% of total energy expenditure. We found that lower AFT predicted weight gain, but only in individuals with BMI 29 kg/m. This indicates that increased insulation associated with adiposity leads to lower cost of metabolizing macronutrients. As individuals gain weight, their energy expenditure increases more than would be expected based on their increased weight. We have found that increased fasting plasma glucose which may be a marker for increase hepatic glucose production explains part of the large than expected increased EE, and predicted less weight gain. Given that energy expenditure and substrate oxidation continue to play an important role in maintaining energy balance, we have continued to investigate factors which explain the inter-individual variance in these measurements. On a cellular level, in adipocytes we found that higher in vitro lipolysis is associated with higher whole body lipid oxidation and less weight gain. We also found that sphingolipid concentrations (which are cell membrane components important in cellular signaling) in skeletal muscle are associated with energy expenditure and weight gain. Furthermore, increased endocannabinoid concentrations in skeletal muscle are also associated with lower energy expenditure. In mediation analysis, the endocannabinoid anandamide appears to be the mediator of sphingomeylins effect on energy expenditure. Genetic factors underlie adiposity and its risk factors. Mutations in the melanocortin 4 receptor gene are associated with increased body mass index and lower 24 hour energy expenditure in humans. Individuals with MC4R mutations have accelerated weight gain in childhood but not in adulthood indicating a more potent effect of this mutation in early life. Furthermore, we found that presence of MC4R mutations predicted development of diabetes in childhood independent of body weight. In an analysis of weight gain trajectories in childhood we have confirmed that children with MC4R mutations cluster into the highest weight gain trajectory. The mechanism by which MC4R leads to hyperphagia and weight gain is not clear. One possible mediator is brain derived neurotrophic factor (BDNF) a downstream effector of MC4R signaling which has been implicated in childhood hyperphagia and weight gain. However, serum BDNF did not differ between individuals with and without MC4R mutations. Recently a more common single nucleotide polymorphism (SNP) has been identified that is associated with increased BMI in Native Americans of southwestern heritage. This common SNP is near the promoter region and is also associated with lower energy expenditure and increased ad libitum food intake. Using a custom genotyping array designed specifically for Native Americans of southwestern heritage, we have also identified a variant in g protein couple receptor (GPR158). GPR158 is involved in neurotransmitter signaling, but we found an association with this variant and lower 24 hour EE and lower resting metabolic rate.